Download Full text in pdf format

Vol. 181: 107-124.1999
MARINE ECOLOGY PROGRESS SERIES
Mar Ecol Prog Ser
Published May 18
Experimental otter trawling on a sandy bottom
ecosystem of the Grand Banks of Newfoundland:
analysis of trawl bycatch and effects on epifauna
Jens Prenal, Peter schwinghamer2,Terence W. Rowell1,Donald C. Gordon ~ r ' , :
Kent D. Gilkinson2,W. Peter Vassl, David L. ~ c ~ e o w n ~
'Department of Fisheries and Oceans, Marine Environmental Sciences Division. Bedford Institute of Oceanography.
PO Box 1006, Dartmouth, Nova Scotia B2Y 4A2, Canada
'Department of Fisheries and Oceans, Northwest Atlantic Fisheries Centre, PO Box 5667, St. John's,
Newfoundland A1C 5 x 1 , Canada
3Department of Fisheries and Oceans, Ocean Sciences Division, Bedford Institute of Oceanography, PO Box 1006, Dartmouth.
Nova Scotia BZY 4A2, Canada
ABSTRACT An experimental study of the effects of otter trawling was conducted in a deep (120 to
146 m) sandy bottom ecosystem of the Grand Banks of Newfoundland from 1993 to 1995. Each year,
three 13 km long corridors were trawled 12 times within 31 to 34 h with an Engel 145 otter trawl
equipped \nth rockhopper foot gear. The width of the disturbance zones created was on the order of
120 to 250 m The total biomass of invertebrate bycatch in the trawl decreased significantly over the 12
sets, even though only a very small proportion of the biomass present was removed and each set did
not pass over exactly the same area of seabed. An influx of scavengmg snow crabs Chlonoecetes opilio
into the trawled corridors was observed after the first 6 sets (approximately 10 to 12 h). Benthic organisms in trawled and nearby reference corridors were sampled with an epibenthic sled. Their biomass
was on average 24 % lower in trawled corridors than in reference corridors. At the species level, this
biomass difference was significant for snow crabs C. opilio, sand dollars Echinarachniusparma, brittle
stars Ophiura sarsi, sea urchins Strongylocentrotuspallidus and soft corals Gersemia sp. The reduced
biomass of epibenthic organisms in trawled corridors is thought to be due to several interacting factors
including direct removal by the trawl, mortality, damage, predation and migration. The homogeneity of
the macro-invertebrate community collected by epibenthic sled was lower in trawled corridors. Sand
dollars, brittle stars and sea urchins demonstrated significant levels of damage from trawling. The
mean individual biomass of epibenthic organisms was lower in trawled corridors suggesting size
specific impacts of trawling, especially for sand dollars. No significant effect of trawling was observed
in the 4 dominant mollusc species captured by the sled (Astarte borealis, Margarites sordidus,
Clinocardium ciliaturn and Cyclocardia novangfiae). This experiment indicates that otter trawling on a
sandy bottom ecosystem can produce detectable changes on both benthic habitat and communities, in
particular a significant reduction in the biomass of large eplbenthic fauna.
K E Y WORDS: Mob~lefishing gear impacts . Otter trawling . Bycatch . Epibenthic organisms - Marine
biodiversity . Grand Banks
INTRODUCTION
Mobile fishing gear is a widespread cause of physical
bsturbance to the global continental shelf benthos (Dayton et al. 1995).Mobile gear such as beam trawls, otter
trawls, scallop rakes and clam dredges inflict damage
'Addressee for correspondence.
E-mail: [email protected]
0 Inter-Research 1999
Resale of full article not permitted
and mortality on benthic invertebrates as well as on target species (e.g.Gruffydd 1972, Caddy 1973, Meyer et
al. 1981, McLoughlin et al. 1991, Andrew & Pepperell
1992, Bergman & Hup 1992, Fonds et al. 1992, Suuronen
et al. 1996).Of the variety of mobile gear types used, the
otter trawl is considered to be one of the more environmentally benign (Hall 1994);nonetheless, it is still capable of inflicting considerable disturbance on benthic
ecosystems (Raloff 1996, Jennings & Kaiser 1998).
Mar Ecol Prog Ser 181: 107-124, 1999
Over the past 50 yr, otter trawls have been the most
commonly used mobile fishing gear on the Atlantic
continental shelf of Canada (Messieh et al. 1991).
Throughout this period, there has been debate concerning the ecological effects of this gea.r. This debate
has recently gained intensity with the collapse of
groundfish stocks in the region in the early 1990s
(Hutchings 1996). Past scientific studies have demonstrated that mobile gear can disturb benthic habitats
and communities, with the discernable effects depending upon the scale of resolution, the kind of gear, the
nature of the substrate and the kind of organisms present. Until recently, it has been difficult to experimentally demonstrate the anticipated impacts of otter
trawling on offshore fishing banks, owing largely to
limitations in experimental design and the technology
required for trawling and sampling within a precisely
defined area on the seabed. Another complicating factor has been that any impacts must be detected against
a background of considerable natural variability in
organism abundance and community structure which
is poorly understood. Nevertheless, studies to date
indicate that larger, longer-lived epibenthic organ-
Fig. 1 Location of the study site on the Grand Banks, Newfoundland. Estimated intensity of seabed disturbance by otter
trawling on the Grand Banks and Labrador SheU in 1990
expressed as the percent of the sampling area swept by a
trawl in 1 yr (P. Schwinghamer & D. W Kulka unpubl. data).
The darkest shade represents a trawling intensity greater
than 8%
isms, especially sedentary and attached forms, can be
adversely affected by mobile fishing gear that comes
into contact with the seabed (Caddy 1973, de Groot
1984, Messieh et al. 1991, Bergman & Hup 1992, Eleftherlou & Robertson 1992, Jones 1992, Thrush et al.
1995, Collie et al. 1997, Jennings & Kaiser 1998).
In 1990, the Department of Fisheries and Oceans
(DFO)initiated a long-term, multi-disciplinary program
to study the impacts of mobile fishing gear on the marine
benthic habitat and communities of Atlantic Canada.
One of the first initiatives was to conduct an experiment
on the effects of otter trawling on the intertidal habitat
and communities in the Minas Basin of the Ba.y of Fundy
(Brylinsky et al. 1994).The observed impacts were minor
but the results should not be extrapolated to offshore
fishing banks because of the major differences in water
depth, sediment type, energy levels and biological communities. Another early initiative was to compile information on the extent and intensity of otter trawling off
eastern Canada. Analysis of archived side-scan sonar
data, collected by the Geological Survey of Canada for
other purposes, indicated that less than 2 % of the available records showed any visible t r a w h g disturbance on
the Scotian Shelf (Jenner et al. 1991) while less than 10 %
showed evidence of trawling disturbance on the Grand
Banks (P. Schwinghamer unpubl. data). Disturbance
from mobile fishing gear was usually found only in areas
of low sediment transport suggesting that effects might
be rapidly obliterated on high energy seabeds and that
side-scan sonar records are inadequate for estimating
the distribution of trawling activity. Krost et al. (1990)
recorded higher frequencies of trawl door tracks with
increasing depth which they interpreted as being
partially the result of decreasing wave energy. Subsequently, a detailed study of the spatial pattern of otter
trawling was conducted for the Grand Banks and
Labrador Shelf using DFO data on commercial fishing
effort for the period of 1980 to 1991 (P. Schwinghamer &
D. W. Kulka unpubl. data). The results showed a very
patchy distribution of trawling effort in this region
(Fig. 1).The most intense effort was limited to a relatively small area (withm which 8 % or more of the seabed
was disturbed in a particular year) while large areas
were untrawled each year.
In 1993, once adequate sampling equipment had been
developed and suitable navigation instrumentation was
available, a major 3 yr experiment on the effects of otter
trawling on a sandy bottom ecosystem was initiated on
the Grand Banks of Newfoundland. This paper presents
an analysis of the otter trawl bycatch data and the effects
of the otter trawling on benthic organisms collected with
an epibenthic sled. No attempt is made to discuss the
effects of trawling on infaunal species since an extensive
series of grab samples was also collected and will be
reported in a future publication..
Prena et al.: Effects of otter trawling on benthic epifauna
METHODS
Study site. The 20 X 20 km study site centred at
47" 10' N, 48" 17'W was established on the northeastern part of the Grand Banks of Newfoundland (Fig. 1).
Until the early 1990s, the fishery in this general region
was for mixed groundfish, mostly American plaice
Hippoglossoides platessoides, with about 15 to 20%
Atlantic cod Gadus morhua. Analysis of fishing effort
records indicated that the study site had not been
fished intensively since the early 1980s (Kulka 1991).
In 1992, the DFO formally closed the site to all fishing
activity for the duration of the experiment. In addition,
as a result of the collapse of the northern cod and other
groundfish stocks in the region, a general moratorium
on groundfish harvesting has been in effect for the
Newfoundland and Labrador continental shelf since
1992. Therefore, we are confident that the only trawling which took place at the study site in recent years
was that conducted in the experiment.
The depth range at the study site is 120 to 146 m.
Surficial sediments are composed of Adolphus sand, a
moderately to well sorted fine- to medium-grained
sand (mean grain size -170 pm) found around the
perimeter of the Grand Banks and on the Flemish Cap
(Fader & Miller 1986).The seabed at this depth is relatively stable with no evidence of wave-induced ripples
on the smooth continuous surface (Barrie et al. 1984).
However, interannual variations in sediment grain size
and acoustic properties were observed during the
experiment, possibly caused by winter storms
(Schwinghamer et al. 1998). Exploratory sampling,
conducted in 1992, demonstrated that the benthic community was species rich (at least 139 macrofaunal species), had a high number of epibenthic species, was
high in biomass (mean of 6 grabs was 650 g wet wt m-')
and had a relatively homogeneous species composition
(Prena et al. 1996). Secondary hard-bottom structures
such as mussel reefs are not present and sessile epifauna1 species (e.g. the soft coral Gersemia sp. and the
bryozoan Myriapora sp.) occur in low densities. The
patchiness of epibenthic megafauna in this region of
the Grand Banks has been described by Schneider et
al. (1987).
Experimental design. Complete details of experimental design are given in Rowel1 et al. (1997) and
briefly summarized here. Based on the experience of
Dutch colleagues investigating the effect of beam
trawls (e.g. Bergman & Hup 1992), it was decided to
conduct a multi-year experiment that included initial
sampling, experimental trawling of known intensity
with commercial fishing gear and post-trawling sampling. Three 13 km long experimental corridors were
established for trawling (Fig. 2). Parallel reference corridors were established 300 m to one side. Each corri-
R
-
T
Comdor A
Fig. 2. Relative position and orientation of Corridors A, B and
C in the study area (20 X 20 km) (not to scale). The central axis
of reference corridors (R) was 300 m to one side of the trawled
corridors (T).Ends of the corridors are approximately 3.5 km
apart
dor was divided into 260 sampling blocks (50 m long
and 200 m wide). Different blocks were selected for
sampling on each cruise using a random procedure
(McKeown & Gordon 1997).The 3 experimental corridors were trawled by the CSS 'Wilfred Templeman', as
described below, in early July for 3 successive years
(1993, 1994 and 1995). Pre- and post-trawl benthic surveying and sampling were conducted each year by the
CSS 'Parizeau'. Benthic surveying and sampling were
also conducted in September 1993, about 10 wk after
the initial trawling disturbance.
Navigation. The 'Parizeau' was positioned with an
accuracy of 3 to 4 m throughout the study using the
Differential Global Positioning System (dGPS).During
the first year of the experiment, the 'Templeman' was
navigated along the trawled corridors using Loran-C
and non-differential GPS. However, dGPS was installed on the 'Templeman' for trawling in 1994 and
1995. The otter trawl, sampling and surveying equipment were positioned relative to the 'Parizeau' by
means of a n ORE International Inc. model 4410C
Trackpoint I1 ultra-short baseline acoustic tracking system. The accuracy of these geo-referenced positions
varied on the order of 4 to 20 m depending upon the
distance between the equipment and the ship. Navigation information (dGPS position, Trackpoint fix, gyro,
log and echo sounder) was merged into a single ASCII
data stream which was then supplied to the ship's
bridge and the scientific laboratories. The software
package AGCNAV was used to graphically display
and record the navigation information. This graphicsbased real-time PC display proved to be the key to the
efficient prosecution of the surveying and sampling
strategies. The tracks of the epibenthic sled while sam-
110
Mar Ecol Prog Ser l
pling on the seabed (approximately 50 m long), determined by Trackpoint, were plotted over the paths of
the otter trawl doors estimated from the navigation
data (McKeown & Gordon 1997). This analysis confirmed that, with just 1 exception, all sled samples in
the trawled corridors were collected entirely within the
area that had been trawled, and that all reference samples were collected well outside the disturbed zone.
Experimental trawling. Experimental trawling was
carried out using an Engel 145 otter trawl with 1250 k g
polyvalent oval otter boards (McCaUum & Walsh 1997).
Approximate door spread was 60 m. This gear was
standard for both commercial fishing and, until 1995,
DFO research surveys in the Newfoundland Region.
Since industry was using rockhopper foot gear, we
used 46 cm (18")rockhopper foot gear rather than the
steel bobbin foot gear used in DFO surveys. The mesh
size (as used in commercial gear) was 180 mm on the
wings and belly of the net and 150 to 130 mm in the
codend. As in DFO surveys, a 30 mm square mesh liner
was installed in the final 9 m of the 18.5 m long codend
in order to capture invertebrates which may be damaged but not retained by commercial gear. Behaviour of the trawl was monitored with a Scanrnar net
mensuration system. Trawling speed was approximately 7 km h-'.
The 3 experimental corridors were trawled within a
5 d period in late June/early July each year (Table 1).
Each year, the corridors were trawled 12 times with
Table 1. Start and end times of the 12 otter trawl sets conducted by the CSS 'Wilfred Templeman' in the 3 experirnental corridors. Also listed are the time intervals between the
end of the last trawl set and epibenthic sled samples. Sled
samples were not collected in Corridor C
Year
Corridor
Trawling
starts
Date
GMT
Trawling
ends
Date
Interval between last trawl
set and sled
GMT
Samples (h)
1993
A
B"
C"
6 Jul 11:30
7 Jul 23:16
8 Jul 18:40
7 Jul 2137
10 Jul 01:20
l 0 Jul 22:38
45-93
10-37
9 Jul 17:31
12 Jul 23:20
11 Jul 0 1 5 9
l 1 Jul 00:24
14 Jul 06:14
12 Jul 0857
12-42
6-13
30 Jun 1 6 ~ 3 4
2 Jul 00:43
3 Jul 10:50
2 Jul 00:07
3 Jul 08:40
4 JuI 1850
9-22
1-5
1994
A
B
C
1995
l
A
-
"Due to equipment problems in 1993, there was a 19 h gap
between Sets 6 and 7 in Corridor B and an 18 h gap between Sets 5 and 6 in. Corridor C
direction alternating on each set. The time required to
trawl the 12 sets in 1 corridor was generally on the
order of 31 to 34 h. Each trawl set was initiated when
the 'Templeman' was about 500 m from the end of the
experimental corridor so that the trawl, which had a
layback on the order of 500 m, would settle into its bottom configuration before entering the corridor. At the
end of each set, the trawl was not retrieved until the
ship was at least 500 m beyond the end of the corridor.
Therefore, the trawl was on the bottom for slightly
more than 13 km for each set. The intention was to
keep all trawling disturbance within 100 m of the centre line. Each year, the 'Parizeau' tracked the position
of the otter trawl for some sets using Trackpoint with a
transponder placed on the head-line. The width of the
disturbance zones was estimated using navigation
data.
Analysis of trawl catch. Trawl catch data were collected in all 3 corridors (Fig. 2). Fish were separated by
species, weighed using girnballed electronic balances
and sized in the ship-board laboratory. The invertebrate bycatch was also sorted by species and weighed.
Subsamples of snow crabs Chionoecetes opilio were
collected for determination of sex, maturity, and shell
condition.
Epibenthic sled sampling. Samples of benthic
organisms were collected with an Aquareve 111 epibenthic sled (Thouzeau & Vine 1991). This gear was
selected because it was available to us and it was felt
that it would collect suitable quantitative samples of
epibenthic organisms at the experimental site if substantially modified. Modifications included decreasing
the width of the sampling blade to 0.34 m, installing a
closing door, widening the runners, and adding wings
to increase stability (Rowel1 et al. 1997).The sled cuts
to a depth of 2 to 3 cm and therefore can collect some
infaunal species as well as epifaunal species. Towing
distance over the seabed, measured by 2 odometer
wheels, averaged 50 m (i.e. each sample represented
about 17 m2).At the end of the tow, the door was closed
electronically and the sled retrieved. Sampling performance was monitored using a video camera mounted
on the sled frame which was directed backward
toward the sled's mouth. Tows of dubious quality (i.e.
lifting off the bottom) were aborted and repeated. A
disadvantage of using an epibenthic sled of this type is
the damage inflicted on captured organisms.
Due to time constraints, sled samples were only collected in Corridors A and B (Fig. 2). Five continuous
sampling blocks were designated for each sled tow
(i.e.250 m along the axis of the corridor).The sled was
equipped with a transponder so that its exact location
on the seabed could be determined using the Trackpoint system. Throughout the entire experiment, 105
sled samples were collected (Table 2 ) . Trawled and
Prena et al.: Effects of otter- trawling on benthic epifauna
2
viduals in the sample. The only exception was that biomass rather than abundance was used for sand dollars
because the weight of fragments could be more accurately determined than the number of individuals they
represented. Chipped mollusc shells and missing or
broken brittle star arms were not considered as damage in our analysis. No damage was assessed for soft
corals Gersemia sp. and snow crabs Chionoecetes
-
opillo.
Table 2. Summary of epibenthic sled samples collected during
the 3 vr e x*~ e r i m e n tSleds
.
were collected either before trawling (B) or after (A). There was no trawling done on Cruise
.
93-029
Cruise
no.
93-021
Tinling
B
A
93-029
94-015
95-013
-
Corridor A
Trawled Reference
Corndor B
Trawled Reterence
2
4
1
1
B
-
-
A
10
10
B
10
10
-
A
111
reference samples were taken from the same box
numbers of the parallel corridors (Fig. 2) (i.e. about
300 m apart), except in 1993. On the initial cruise (93021), 2 sled samples were collected from neighbouring
blocks in each of the trawled and reference corridors
before trawling, and 4 samples were collected in other
blocks of the trawled corridors after trawling. On the
second cruise (93-029),only half the intended samples
were collected because of bad weather. As a conseqence of incomplete experimental blocks, these data
are not included here. Based on an evaluation of the
1993 data, the number of sled samples was increased
to 10 in each reference and trawled corridor during
1994 and 1995. Unfortunately, time constraints made it
impossible to collect sled samples from the trawled corridors before retrawling in 1994 and 1995. Therefore,
the only opportunity to collect sled samples in the
trawled corridors immediately before and after trawling occurred m 1993. The resulting sled data set provides no opportunity to examine the possible recovery
of epibenthic organisms in the trawled corridors before
retrawling. The interval between the completion of
experimental trawling and the collection of sled samples in trawled corridors was quite variable and an
attempt was made in 1994 and 1995 to reduce it as
much as possible (Table l ) .It was planned to sample
the reference corridors before trawling. However, due
to operational constraints, in 1994 reference samples
had to be collected after trawling (Table 2).
The sled catch was washed with seawater, under low
pressure, over a 1 mm mesh screen. Common invertebrates were sorted by species and damage categories,
counted and weighed on-board using suspended
spring balances. Less common species were frozen for
later processing ashore. Sled catch data were converted to units per m2 of seabed using odometer data.
Physical damage to echinoderms (e.g.broken test) and
molluscs (e.g. broken shell) was assessed during sorting and the number of damaged organisms was
expressed as a percentage of the total number of indi-
Statistics. Invertebrate bycatch b~omassdata were
analysed using a repeated measures ANOVA with corridors (A, B, C) as experimental subjects, nested in
years (1993, 1994 and 1995),and bycatch in the 12 consecutive trawl sets as repeated measures. Biomass data
were log-transformed in order to satisfy assumptions of
the ANOVA. Where it was necessary to use ranked
data, ranks were re-expressed as folded logs according
to Mosteller & Tukey (1977) for parametric tests of significance. Statistical analyses were performed with
Minitab Version l l (Minitab 1996).
In processing the epibenthic sled data, a 3-way
ANOVA for aggregate data and a 3-way MANOVA for
dominant species were used to determine the statistical significance of the 3 main factors incorporated in
the experimental design; namely trawling (comparing
reference and trawled corridors), corridor (comparing
A and B) and year (comparing 1993, 1994 and 1995).
For statistical analysis, experimental units were
defined as sampling blocks (i.e. stations) within the 2
corridors that were sampled. A paired test design gave
very similar results and is not presented. MANOVA of
acceptable fit (criterion F,,,,,) could be computed for
mean individual biomass and physical damage only
when raw data were ranked or re-expressed by principal components. Both versions yielded similar results.
Here we present MANOVA of those data sets based on
principal components. For this reason the ANOVA of
Table 7 was computed for each species separately and
is not based on the MANOVA shown in Table 6.
Assumptions of ANOVA could usually be met by
applying log or root-transformations. Where it was
necessary to use ranked data, ranks were re-expressed
as folded logs according to Mosteller & Tukey (1977)
for parametric tests of significance. The nominal significance level of 5 % was Bonferroni adjusted (by dividing through the number of univariate tests) to control
for compounding of the experimental error rate.
Hon~ogeneityof the benthic community in trawled
and reference corridors was examined by regression of
mean and standard deviation of the abundance for all
species within each combination of main effects (as
described above for ANOVA). Warwick & Clarke
(1993) showed that means and standard deviations can
be described by linear relationships after transformation which can b e used to compare the relative degree
112
Mar Ecol Prog Ser 181: 107-124, 1999
of variability between samples, or as a n index of
aggregation. An increase of the standard deviation relative to the mean, measured by the slope of the regression, is interpreted as a decrease in homogeneity of the
species community (or increase in aggregation). A
sixth root transformation was applied to reduce the
strong influence of the dominant sand dollars and brittle stars on the slope and to obtain a balanced distribution of data points along the regression line. Statistical
analyses were performed using Systat 5.2 (Systat
1992).
1997). It appears that this offset was due to slight differences in the length of the warps. The estimated
widths of the disturbance zones for Corridors A and B
are given in Table 3. The mean width was greatest in
1993, In excess of the intended 200 m but was reduced
substantially in 1994 and 1995 to approximately 120 m
or less.
Trawl bycatch
Snow crabs Chionoecetes opilio, basket stars Gorgonocephalus arcticus and sea urchins Strongylocentrotus pallidus dominated the trawl invertebrate by-
RESULTS
Distribution of trawling disturbance
Navigation data were not logged for the 'Templeman' d u n n g trawling in 1993, so it was not possible to
plot the exact course relative to the intended line.
However, the Trackpoint data on trawl position indicated that the vessel wandered off course at times as
much as several hundred meters. Navigation data
were logged in 1994 and 1995 when the 'Templeman'
was equipped with dGPS. In 1994, the vessel stayed
very close to the intended line, usually within 20 m . In
1995, the vessel deliberately steered a sinusoidal
course in order to distribute the trawling disturbance
more widely across the trawled corridor.
Scanmar data indicated that the trawl, once settled
into its fishing configuration, had a wing spread (i.e.
net opening) of 20 2 m and a door spread (i.e. distance between the otter boards) of 60 k 5 m on all sets.
Trackpoint data indicated that, while fishing, the trawl
was consistently set to port relative to the ship position
on the order of 20 to 35 m regardless of the direction
steamed along the corridor (McKeo~vn& Gordon
catch and were consistently caught in large enough
quantities to allow analysis of the relationship of
bycatch wlth set number. Soft corals Gersemia sp.,
whelks Buccinum sp. and hermit crabs Pagurus sp.
were commonly encountered in the small amount of
other material that made up the bycatch. Iceland scallops Chlamys islandica and a few other bivalve molluscs were occasionally captured by the trawl.
The invertebrate bycatch biomass (both total and
major components), averaged for all 3 corridors and 3
years, is presented by consecutive trawl set in Fig. 3A.
The mean biomass was on the order of 10 kg (wet
20
-
I ISnow Crab
IBasket Star
*
15
.
Q)
A
ISea Urch~n
o
Others
10
m
Y
5
0
Table 3 E s t m a t e d width of the disturbance zones created by
otter trawling as estimated from the navigation data (McKeown & Gordon 1997) Indiv~dualestimates ( N ) were made by
measunng the dtstance between the outermost door tracks In
those boxes where samples were collected. Door track postt ~ o n swere estimated uslng either the position of the headline
as determined by Trackpoint, if available, or vessel position
L
m
SD
Mm
(m)
(m)
A
B
264
204
1994
A
B
1995
A
B
Year
1993
Corndor
N
(m)
Max
(m1
55
102
160
80
360
420
34
33
81
11.9
10
15
75
90
125
150
30
30
121
112
9
16
110
90
140
160
19
20
Mean
05
m
(I)
0
1
2
3
4
5
6
7
6
9 1 0 1 1 1 2
Consecut~veTrawl Set
Fig 3 [A) Mean and (B) coeffic~entof vanatlon of lnvertebrate bycatch b ~ o m a s splotted agalnst otter trawl set (1 to 12)
Each trawl set was 13 km long T m e needed to conduct the 12
trawl sets was 31 to 34 h Average of 3 corridors (A, B and C)
and 3 years (1993-95)
Prena et al.. Effects of otter trawling on benthic epifauna
weight) for a 13 km set. There was a trend of decreaslng bycatch in the first 6 sets, due primarily to a decline
in snow crabs, with relatively constant levels thereafter. The biomass of basket stars tended to be slightly
lower in later tows but there was no clear trend in sea
urchin biomass. The coefficient of variation (standard
deviatiodmean) of snow crab, basket star and sea
urchin bycatch increased markedly on the first 3 sets
(Fig. 3B), suggesting increasing patchiness. A regression analysis of ranked invertebrate bycatch against
trawl set, while highly variable, indicated a significant
decrease in biomass with increasing set (p < 0.05)
(F1g. 4).
It appears as if some of the non-linearity in bycatch
decline with increasing trawl set (Fig. 3A) may be due
to a behavioural response of snow crabs which are
highly mobile. Ranked numerical abundance of adult
and immediately pre-adult male snow crabs (carapace
width > 60 mm), which made u p more than 70 % of the
crab bycatch biomass, is plotted against set number in
Fig. 5. Biomass decreased markedly during the first 6
sets (approximately 15 to 17 h). Thereafter, biomass
increased again but never reached initial levels.
The fish catch was extremely low and averaged on
the order of 18 kg (wet weight) for a 13 km set (Fig. 6).
It was dominated by American plaice Hippoglossoides
platessoides, but thorny skate Raja radiata were also
abundant. Capelin Mallotus villosus were caught primarily during the daytime. Arctic cod Boreogadus
saida and sand lance Amrnodytes dubius were caught
in small quantities. Individuals of other fish species
common to this area of the Grand Banks were occasionally captured. Very few Atlantic cod Gadus
morhua were caught over the 3 yr experimental
period. There was no indication of any trend in the fish
catch with increasing set number or of differences
between the 3 corridors.
ANOVA of total invertebrate bycatch biomass
demonstrated that the effects of year (1993, 1994,
1995),corridor (A, B, C nested in year) and consecutive
trawl set (1 to 12, repeated measure) were statistically
significant ( p < 0.001). The drop each year in the mean
biomass of snow crabs, basket stars, sea urchins, other
invertebrates, and fish collected by the otter trawl is
illustrated in Fig. 6.
l
1
1
1
2
1
1
1
1
6
4
1
1
1
1
1
12
10
8
Consecutive Trawl Set
Fig. 4 . Relationship of total invertebrate bycatch biomass
(folded log ranks) to otter trawl set (1 to 1 2 ) . Data from 3 corridors (A, B and C) and 3 years (1993-95) ( p c 0.05). CI: confidence interval, PI: prediction interval
- v.--v...
4 Median Rank
--V-.10 Percentile
-
'A...Y
1
1
1
2
1
1
1
4
1
6
1
8
1
1
1
10
1
12
Consecutive Trawl Set
Fig. 5. Relationship of male snow crab (>60 mm carapace
width) biomass to otter trawl set (1to 12).Data are ranked relative abundance (Rank 12 =highest per 12 sets). Average of 3
corridors (A, B and C) and 2 years (1994-95). Time needed to
conduct the 12 trawl sets was 31 to 34 h
Fish
Snow Crabs
F
Basket
Stars
Sea Urch~ns
Macrofaunal community sampled by the epibenthic
sled
General description of epifauna community
The total biomass of organisms collected by the
epibenthic sled along the reference corridors ranged
between 317 and 475 g (wet wt) m-2 (Table 4). Nine
Kg Bycatch
per Set
0
1
4
9
16
25
Fig. 6. Otter trawl fish catch and invertebrate bycatch by year
(1993-95) expressed as wet weight biomass per trawl set.
Average of all sets and 3 corridors (A, B and C)
114
Mar Ecol Prog Ser 181: 107-124, 1999
Table 4. Mean and standard deviation (in brackets) of wet weight (g m-') for the 9 dominant species and the total sample (all
species) collected by epibenthic sled. T, trawled; R: reference. E: Echinodermata, A: Arthropoda; M: Molluscs; C: Coelenterata
Treatment
Echinarachruus parma ( E )
Ophiura sarsi (E)
Strongylocentrotus pallidus (E)
Astarte boreahs (M)
Chionoecetes opiho ( A )
Gersema sp. (C)
Margarites sordidus ( M )
Clinocardium ciliaturn (M)
Cyclocardia novangliae (M)
Total sample (all species)
1993
Corridor A Corridor B
T
R
T
R
T
R
T
R
T
R
T
R
T
R
T
R
T
R
221 (27)
230 (82)
T
R
316 (19)
377 (75)
46 (14)
75 (16)
21 (10)
39 (10)
9.6 (9.6)
5.8 (4.6)
0.3 (0.2)
4.9 (6.9)
1.7 (0.8)
1.9 (0.6)
0.5 (0.2)
0.8 (0.2)
0.2 (0.2)
0.0 (0.0)
0.5 (0.2)
0.5 (0.4)
1994
Corridor A Corridor B
168 (73)
296 (23)
68 (9)
71 (11)
37 (19)
41 (10)
9.9 (13.9)
7 9 (1.2)
8 2 (13.4)
4.9 (6.9)
0.6 (1.0)
2.6 (1.2)
1.6 (1.8)
0.9 (0.4)
1.0 (1.3)
0.9 (0.6)
0.1 (0 1)
0 7 (0.5)
305 (62)
433 (42)
dominant species were present in more than 90% of
the 96 samples included in the data analysis and made
up 95 to 98% of the total biomass. They are, in
decreasing order of mean biomass, the sand dollar
Echinarachnius parma, the brittle star Ophiura sarsi,
the sea urchin Strongylocentrotus pallidus, the snow
crab Chionoecetes opilio, the mollusc Astarte borealis,
the soft coral Gersemia sp. and the molluscs Margantes sordidus, CLinocardium ciliatum and Cyclocardia novangliae. The 3 most dominant species were
echinoderms. The polychaete Nothria conchylega was
also abundant in all sled samples but not processed
because of the extensive sorting time required. The
total number of species collected in all sled samples
was 115 (list available on request). Since the sled blade
cuts to a depth of several centimeters into the sandy
sediment found at the study site, some infaunal organisms were collected (e.g. certain molluscs). Only sea
urchins and snow crabs were commonly caught by
both the otter trawl and sled.
Biomass
222 (48)
243 (70)
36 (14)
56 (21)
13 (8)
24 (10)
10.4 (7.2)
9.6 (5.2)
2.1 (4.8)
3.0 (2.6)
1.6 (1.4)
3.3 (2.1)
0.5 (0.4)
0.5 (0.3)
0.4 (0.2)
0.6 (0.5)
0.6 (0.4)
1.0 (0.8)
162 (69)
181 (54)
76 (27)
73 (21)
42 (24)
35 (13)
3.7 (2.4)
4.0 (3.4)
2.8 (3.7)
5.7 (9.4)
2.0 (1.3)
3.3 (2.8)
0.7 (0.4)
0.5 (0.3)
0.7 (0.7)
0.4 (0.4)
0 2 (0.2)
0.3 (0 3)
218 (49)
180 (69)
321 (65)
275 (68)
41 (15)
40 (39)
81 (23)
88 (22)
34 (20)
23 (19)
35 (9)
34 (10)
10.4 (4.4)
3.8 (2.8)
6.0 (7.1)
7.2 (7.2)
8.5 (8.2)
4.3 (6.0)
13.7 (15.9) 11.1 (9.8)
1.6 ( 1 . 0
2.2 (2.4)
2.0 (1.3)
2.2 (1.4)
0.6 (0.5)
0.4 (0.3)
0.7 (0.4)
0.5 (0.3)
0.4 (0.7)
1.3 (0.6)
0.4 (0.5)
0.7 (1.0)
0.5 (0.3)
0.3 (0.4)
0.7 (0.6)
0.3 (0.4)
291 (55)
352 (48)
303 (72)
317 (50)
327 (59)
475 (78)
261 (76)
426 (61)
initial disturbance (Table 2 ) . In all 3 years, the biomass
in trawled corridors was lower than in reference corridors. Biomass in reference corridors showed considerable interannual variability, while biomass in the
trawled corridors stayed relatively constant with time
despite repeated trawling. The greatest difference in
total biomass between reference and trawled corridors
occurred in the final year of the experiment (1995).The
least difference occurred in 1994 when the reference
corridors were sampled after trawling (Table 2) and
REFERENCE
Corrldor A
Corridor B
250
t
TRAWLED
o Corr~dorA
Corr~dorB
1993
The mean total biomass of organisms sampled by the
epibenthic sled in reference and trawled corridors over
the entire experiment is presented in Table 4 and
Fig. 7. The reference samples in 1993 include the sled
samples taken from the trawled corridors before the
1995
Corridor A Corridor B
1994
1995
Year
Fig. 7. Total biomass of invertebrates collected by the epibenthic sled in trawled and reference corridors ( A and B only)
each year of the experiment Shaded a.reas represent 95%
conf~denceintervals
Prena et a1 : Effects of otter trawling on benthic epifauna
Table 5. Three-way ANOVA of total biomass and mean individual biomass of organisms collected in the epibenthic sled
for effects of corridor (A and B), trawling (trawled and reference) and year (1993, 1994 and 1995). Sources of variance in
bold are statistically significant
Table 6. Three-way MANOVA of biomass, mean individual
biomass and physical damage (without soft corals and snow
crabs) of the 9 dominant species collected in the epibenthic
sled on effects of corridor, trawling and year. Sources of variance in bold are statistically significant
Variable
Source
Variable
Source
Total
biomass
Corridor
Trawling
Year
Error
Biomass
(9 species)
Corridor
Trawling
Year
Error
Mean
individual
biomass
Corridor
Trawling
Year
Error
Mean
individual
biomass
(9 species)
Corridor
Trawling
Year
Error
Physical
damage
(7 species)
Corridor
Trawling
Year
Error
migration of scavengers in response to trawling may
have reduced the biomass in reference corridors. The
3-way ANOVA indicated that the differences in total
biomass between the trawled and reference corridors
and year were statistically significant and that trawling
accounted for most of the variance in the data
(Table 5 ) . Differences in total biomass between Corridors A and B were not significant. Overall, the total
biomass in the trawled corridors was on the order of
24% less than in the reference corridors. A 3-way
MANOVA indicates that the effects of trawling, year
and corridor are significant when the biomass of the 9
dominant species is tested (Table 6). Again, the greatest variance is associated with trawling.
A 3-way ANOVA indicates that the biomass of 5 of
the 9 dominant species was significantly lower in the
trawled corridors than in the reference corridors
(Table 7). Sand dollars Echinarachniusparrna and brittle stars Ophiura sarsi had the greatest probability of
being affected followed by soft corals Gersemia sp.,
snow crabs Chionoecetes opilio and sea urchins
Strongylocentrotus pallidus. There were trends of
lower biomass of the mollusc Cyclocardia novangliae
and increasing biomass of the n~olluscClinocardium
ciliatum in trawled corridors (Table 4 ) but these were
not statistically significant by our criterion of significance (p < 0.017) (Table 7). The ANOVA indicated that
the biomass of the molluscs Astarte borealis and Marwrites sordidus had the least probability of being
affected by trawling.
Mean individual biomass
Analyses of the abundance data gave similar results
as the biomass data and are therefore not presented.
However, the potential effects of trawling on organism
size was examined by dividing biomass by abundance
115
Wilks' A
F
to obtain mean individual biomass (Table 8). With the
exception of Corridor A in 1994, the mean individual
biomass for all species captured was less in the trawled
corridors. The difference was statistically significant
by 3-way ANOVA (Table 5). This difference in size
was not significantly related to trawling when the
mean individual biomass of the 9 dominant species
was tested by the 3-way MANOVA (Table 6).This test
also demonstrated a significant effect of both corridor
(smaller in Corridor B) and year. The 3-way ANOVA of
the mean individual biomass of the separate 9 dominant species indicates that only the sand dollar Echinarachnius parma was significantly affected by trawling (Table ?). A similar trend of smaller mean individual biomass in trawled corridors was apparent in the
mollusc Cyclocardia novangliae but was not significant (p = 0.105).
Damage
Physical damage was quite variable by species
(Table 9 ) . Approximately half of some species were
damaged (e.g.the molluscs Astarte borealis and Clinocardium ciliaturn) while others suffered lower levels of
damage (e.g. the brittle star Ophiura sarsi and the
mollusc Margarites sordidus). The high percentage of
damage for some species collected from reference corridors indicates that considerable damage was inflicted
by the sled during sampling. For example, sampling
apparently damaged 15 to 60 % of the mollusc A. borealis, 8 to 23% of the sand dollars Echinarachnius
parma and 1 to 5 % of the sea urchins Strongylocentrotus pallidus. However, there was a general trend of
Mar Ecol Prog Ser 181: 107- 124, 1999
116
Table 7. Specific effect of trawling on biomass, mean individual biomass and physical damage of dominant benthic species collected by epibenthic sled. Three-way ANOVA with the effects trdwling, corridor and year. Biomass model based on MANOVA
shown in Table 6. Sources of vdriance in bold are statistically significant. Criterion of significance is 0.017 after Bonferroni adjustment. E: Echinodermata; A: Arthropoda; M: Mollusca; C: Coelenteratd
I
Variable
Species
Biomass
Astarte borealis (M)
Clinocardium cilia turn (31)
Cyclocardia novangliae (hf)
Chionoeceles opilio (A)
Echinarachnius parma (E)
Gersemia sp. (C)
Margantes sordidus (hl)
Ophiura sarsi (E)
Strongylocentrotus pallidus (E)
Mean
indiv~dual
biomass
Astarte borealis (M)
Clinocardium ciliatum ( M )
Cyclocardia novangliae" (M)
Chionoecetes opilio (A)
Echinarachnius parrna (E)
Gersemia sp. (C)
Margantes sordidus (M)
Ophiura sarsiVE)
Strongylocentrotus pallldusd (E)
Physical
damage
Astarte borealisd (M)
Echinarachnius parrna (E)
Ophiura sarsi (E)
Strongylocentrotus pallidus" (E)
MS effect
MS error
df error
0.493
3.810
8.144
72 753
2.785
0.193
0.207
1 600
90
91
88
90
F
0.177
19.719
39.250
45.457
I
p
0.675
0.000
0.000
0.000
I "Data ranked
I
Table 8. Mean and standard deviation (in parentheses) of mean individual biomass (g wet weight) for the 9 dominant species and
the total sample (all species) collected by epibenthic sled Undamaged specimens of Echinarachnius parma were considered
only T: trawled; R: reference. E: Echinodermata; A: Arthropoda; M: Mollusca, C: Coelenterata
Treatment
Corridor A
Chionoecetes opilio ( A )
Strongylocen trotus paliidus (E)
Astarte borealis (h?)
Ech~narachniusparma (E)
Ophjum sarsi (E)
Gersemia sp. (C)
Clinocardium ciliatum (M)
blaryarites sordidus (M)
Cyclocardia novangliae (hi)
Total sample (all s p e c ~ e s )
1993
Corridor B
1994
Corndor A Corridor B
T
R
T
R
T
R
T
R
T
R
T
R
T
R
T
R
T
R
15( 1
9.4 (9.8)
16.6 (2.0)
13.8 (3.8)
8.6 ( 1 . 1
11.5 (4.8)
4.5 (0.5)
5 . 1 (0.6)
3.7 (0.5)
3.6 (0.6)
2.0 (1.6)
1.4 (0.9)
1.8 (0.4)
0.6 (0.3)
1.4 (0.3)
1.7 (0.4)
1.1 (0.3)
1 2 (0.1
9.8 (9.6)
7.3 (8.1)
9.5 (0.6)
12.4 (1.8)
8.2 (4.4)
7.0 (0.5)
4 2 (0.7)
5 4 (0.5)
2.5 (0.9)
3.6 (0.2)
0.8 (0.3)
5.7 (2.1)
1 9 (1.9)
3 6 (3.1)
2.9 (2.8)
1.3 (0.2)
1.1 (0.3)
1.7 (0.4)
8.9 (18.4)
4.9 (3.5)
15.4 (4.2)
14.7 (3.2)
8.1 (2.3)
7.0 (1.8)
4.7 (1.3)
4.4 (1.2)
4 9 (0.7)
4.8 (0.6)
1 9 (1.8)
1.4 (0.9)
2 1 (1.8)
1 5 (0.9)
1.2 (0.4)
1.2 (0.4)
1.2 ( l )
1.4 (0.2)
19.6 (34 3)
11 4 (14 6)
15.3 (3.0)
14.1 (1.7)
5.5 (2.8)
5.4 (3.2)
4 1 (1.7)
4 4 (1.5)
3 5 (1 4)
3.5 (1.2)
3.9 (2.8)
3.9 (3.2)
1 8 (1 0)
1.6 (1 5)
1.3 (0 4)
1.4 (0.5)
0.8 (0.4)
0.9 (0.3)
T
R
5.3 (0.2)
5.5 (0.51
3.9 (1.0)
5.4 (0.5)
5.2 (1.2)
4.9 (1.0)
4.5 (1.9)
4.6 ( l 5 )
1995
Corridor A Corridor B
55.2 (37 6 ) 16.9 (18.2)
33 3 (47.1) 57.1. (34.0)
16.2 (6 1.)
15.3 (2 5)
19.5 (3.2)
16.9 (3.7)
8.9 (2.6)
7.6 (2.1)
8.5 (3.1)
6.4 (3.2)
3.8 (0 6)
3.9 (1.3)
5.8(1.4)
5.9 (1.41
5.7 (1.2)
4.5 (2 2)
5.9 (3.8)
6.6 (7.3)
1.9 ( 1 . 0
2.5 (1.9)
3.6 (2.7)
1.5 (0.8)
2.4 (2.9)
4.1 (5.0)
1 6 (10)
1.8 (0.6)
1.5 (0.4)
1.2 (0.3)
1.1 (0.3)
1.4 (0.3)
1.1 (0.4)
1.3 (0.3)
1.1 (0.5)
1.4 (0.2)
4.9 (0 7)
6.0 (1 1)
4.5 (1.6)
5.6 (1.9)
117
Prena et al.: Effects of otter tra\vling on b e n t h ~ cepifauna
Table 9. Mean and standard devlation (In parentheses) of the percentage of 7 dominant species collected by epibenthic sled that
were damaged. T- trawled, R: reference E: Echinodermata, IM: Mollusca. Soft corals and snow crabs could not be assessed for
damage
Treatment
Corridor A
1993
Corridor B
1994
Corridor A Corridor B
Astal te borealis ( M )
T
R
72 8 (20 2) 40 1 (10 8 )
14 8 (11 1 ) 3 8 3 (4 7)
55 5 (12 0 )
58 0 (21 7 )
Clinocardium c~liatum(M)
T
R
20 9 (13.4) 45 1 (26 2)
10 6 (17 4) 42.2 (7.2)
44 0 (28.6) 59.4 (32.2)
51.1 (29.8) 59.1 (35.3)
Cyclocardja novangllae ( M )
T
R
41 7 (22 0)
0
16 7 (28 9)
9.7 (11.6)
17 9 (11.6) 15.9 (32.0)
26.4 (27.8) 38.5 (38.9)
Echinarachnius parma (E)
T
R
21 9 (4.1)
22.8 (6.2)
18.6 (9.0)
8.4 (5.1)
14.8 ( 5 1)
15.5 (3.9)
27.6 (9.5)
22.8 (6.8)
Strongylocentrotus palljdus (E)
T
R
7 . 0 (8.4)
0.5 (1.0)
1.8 ( 1 6 )
2.0 (1.5)
17.9 (25.8)
4 . 6 (4.3)
12.1 (12.0)
0.5 ( 1 2)
Margarjtes sordidus ( M )
T
R
2 3 (0.6)
2.5 (5.0)
1 5 (1.0)
2.5 (5.0)
8 2 (16.5)
0
1.2 (4 0 )
0.7 (2 3)
1995
Corridor A Corndor B
47 9 (26 0 )
54 6 (25 2)
Ophjura sarsl (E)
greater damage in trawled corridors compared to reference corridors for all species (Table 9). The 3-way
MANOVA indicated that there were significant effects
of both trawling and year on damage but not of corndor (Table 6 ) . Analysis of the effect of trawling on individual species indicated that only the 3 echinoderm
species had significantly higher damage in the trawled
corndors (Table ?), with the greatest probability of
impact on the sea urchins (on the order of 1 0 % dama g e ) . Trawling did not have a significant effect on the
damage of A. borealis. Because of the relatively large
proportion of samples without damaged specimens, it
was not possible to perform a parametric test of significance on the influence of trawling on the n~olluscs
M. sordidus, C. ciliaturn, and Cyclocardia novangliae.
However, all dominant species could be used in the
above mentioned MANOVA after principal
component analysis.
Table 10 Regression of root-transformed means and standard deviations
for the abundance of all species collected by epibenthic sled in each
combination of the main effects of trawling, corrldor and year. N: number of taxa; SE: standard error
N
R2
Slope
(SE)
Intercept
(SE)
Reference
Trawled
Reference"
Trawledd
Reference
Trawled
45
36
44
35
51
48
0.968
0.921
0.952
0.944
0 946
0.929
0.729 (0.020)
0 640 (0 032)
0 724 (0.025)
0.740 (0.031)
0.569 (0.019)
0.799 (0.033)
0 210 (0.016)
0.261 (0.025)
0.213 (0.019)
0 193 (0.023)
0.306 (0.015)
0.146 (0.025)
Reference
Trawled
Reference
Trawled
66
60
70
70
0.972
0.975
0.956
0.969
0.662 (0.014)
0.736 (0.015)
0.751 (0.020)
0.803 (0.017)
0.253 (0.010)
0.213 (0.011)
0.202 (0.014)
0.168 (0.012)
Reference
Trawled
Reference
Trawled
60
71
68
64
0.956
0.933
0.943
0.956
0 691 (0 019)
0.735 (0.024)
0.761 (0 023)
0.818 (0 022)
0.241 (0.014)
0.208 (0.016)
0.200 (0.016)
0.164 (0.016)
Corrldor Treatment
-1993
A
B
1994
A
B
1995
A
B
"Sand dollars excluded
Species homogeneity
A linear relationship between the mean
abundance and standard deviation of all species in each treatment group (year, corridor
and trawling) was established by applying
root-transformation (Fig. 8 ) . Since sand dollars Echinarachnius parma and brittle stars
Ophiura sarsi were very dominant, a sixth
root transformation was used to achieve a
more even distribution of the data points
along the regression lines. The coefficient of
determination was >0.92 for all regressions
(Table 10). The slopes of the regressions of
trawled corridors were greater than those of
reference corridors, except for Corridor A in
1993 (Fig. 8 ) . However, elimination of the
sand dollar data in this instance improved
the coefficient of determination for the
trawled corridor and reversed the relationship of the regression coefficients between
trawled a n d untrawled areas (Table 10).With
118
Mar Ecol Prog Ser 181: 107-124, 1999
Corridor A
1993
-
Trawled
Corridor B
Sixth Root of Mean Abundance per Meter Square
Fig. 8. Linear regression of root-transformed standard deviations and means for the abundance of all species sampled by the
epibenthic sled in trawled and reference corridors (A and B) in 1993-95. Regression parameters are given in Table 10
this restriction, there was a general increase in the
standard deviations relative to the means in the
trawled corridors, indicating decreased homogeneity
and increased aggregation in the benthic community.
DISCUSSION
The intensity of trawling applied in this experiment,
12 sets along the same line in each of 3 consecutive
years, was intentionally designed to be high. Since the
width of the disturbance zone '~vlthinthe trawled corridors (Table 3) was on the order of 2 to 4 times greater
than the spread of the otter trawl doors (about 60 m),
the same area of seabed was not disturbed on all sets
and the average trawling intensity for a given location
on the seabed was probably on the order of 3 to 6 sets
per year. This trawling intensity appears to be higher
than has occurred on the Grand Banks in recent years
as estimated from effort data (Fig. 1).As noted above,
only a small area of the Grand Banks was fished in
1990 at a level where 8% or more of the seabed had
been trawled. This is the equivalent of 1 trawling disturbance every 12 yr. However, commercial fishing
effort can be more intense in other nearby regions. For
example, Auster et al. (1996), using effort data collected between 1976 and 1992, estimate that on average the U.S. sector of the Gulf of Maine is trawled once
a year while the U.S. sector of Georges Bank is trawled
3 to 4 times a year. Because of the pronounced patchiness of trawling effort (Fig. l ) , targeted fishing
grounds can receive an even higher level of disturbance, since it is common practice to fish a productive
area heavily until catch declines to the point where further effort is not cost effective. Therefore, the intensity
of trawling used in our experiment is not unrealistic.
The effects of the otter trawling on the physical habitat observed in this expenrnent have been presented
by Schwinghamer et al. (1998) but are briefly summarized here. Trawling had no observed effect on sedi-
Prena et al.: Effects of otter trawling on benthic epifauna
ment grain size. There was an immediate effect on sediment surface topography due to the berms and furrows created by the trawl doors which were readily
detected by sidescan sonar and at times visible in
video imagery. In most instances, they disappeared
within 1 yr. RoxAnnTMdata also indicated changes in
sediment surface topography with increasing trawl set.
Changes caused by trawling in the acoustical properties of the upper 4.5 cm of sediment were detected by
DRUMSrMand suggest decreased habitat complexity
through destruction of biogenic structures such as
tubes and burrows (Schwinghamer et al. 1996). High
resolution colour video observations indicated that
trawling reduces biogenic sediment structures and the
abundance of flocculated organic matter on the sediment surface. However, these visual effects disappeared in less than a year. In summary, these observations indicate that the physical habitat at the study site
recovered from the experimental trawling disturbance
within approximately 1 yr. It is important to note that
there were significant interannual changes in both
sediment grain size and acoustic properties that could
not be attributed to trawling. Possible natural sources
of this variability include bioturbation and winter
storms.
As expected, the epibenthic sled catch was quite different from the invertebrate bycatch of the otter trawl.
Only the snow crab Chionoecetes opilio and the sea
urchin Strongylocentrotus pallidus were common in
both. The third dominant invertebrate in the trawl
bycatch, the basket star Gorgonocephalus arcticus,
was only occasionally captured by the sled. Assuming
a n average invertebrate bycatch of 10 k g tow-' (Fig. 3),
taking into account the length of the tow (13 km) and
the width of the area sampled (20 m), the otter trawl
captured an average invertebrate biomass on the order
of 38 mg m-2, w h ~ c h1s only 0.01 % of that captured by
the sled (Table 4 ) . A more conservative estimate of
0.5 % is obtained using only the 3 dominant species of
the invertebrate bycatch and a sampling width of 5 m
(i.e. codend with smaller mesh). These results indicate
that the otter trawl is very inefficient in catching
epibenthic species, presumably due to its larger mesh
size and its practice of travelling just above the sediment surface rather than cutting through it as the sled
does. Different modes of capture (e.g. turbulent displacement into the otter trawl versus scooping up into
the sled) also may have contributed to the observed
differences in catches.
Even though the trawl is very inefficient in sampling
epibenthic organisms and did not pass over exactly the
same area of seabed on each set, there was a significant decrease of invertebrate bycatch in the otter trawl
over the 12 consecutive sets (Figs. 3 & 4). It is questionable whether this decrease was caused entirely by the
119
direct removal of epibenthic organisms by the trawl.
Another contributing factor could be that, once disturbed or damaged, epibenthic organisms are more
difficult to capture in the trawl during later sets. The
biological damage inflicted by the otter trawl was
much greater than indicated by the biomass of invertebrate bycatch. The remains of dead organisms were
commonly seen fouling the wings and belly of the net.
In addition, large numbers of damaged organisms left
behind on the seabed were detected in the sled samples collected soon after trawling (Table 9). The 3 species significantly damaged were sand dollars Echinarachnius parma, brittle stars Ophiura sarsi and sea
urchins Strongylocentrotus pallidus, all echinoderms
(Table 7). These estimates of damage may be low
because of predation by scavengers during the interval
between trawling and sampling which in some
instances was greater than several days (Table 1).
Video imagery showed that basket stars Gorgonocephalus arcticus were coiled into tight balls after passage of the otter trawl but many seemed to recover
within a day (Schwinghamer et al. 1998). The relative
consistency of basket star and sea urchin biomass in
the invertebrate bycatch (Fig. 3) reflects their low
catchability by the trawl and the likely possibility that
later sets passed over substantial areas of unsampled
bottom because of variability in the path of the trawl.
The total biomass of organisms captured by the
epibenthic sled was consistently lower in the trawled
corridors in each of the 3 years of the experiment
(Table 4, Fig. 7). The decrease averaged 24 % and was
statistically significant (Tables 5 & 6). Analysis at the
species level indicated that the decreases were due
primarily to reductions in the biomass of sand dollars
Echinarachnius parma, brittle stars Ophiura sarsi, soft
corals Gersemia sp., sea urchins Strongylocentrotus
pallidus, and snow crabs Chionoecetes opilio (Table 7 ) .
Because of its relative abundance in the invertebrate
bycatch of the otter trawl (Fig. 3), it is possible that the
biomass of the basket star Gorgonocephalus arcticus
was also reduced by trawling. However, we were
unable to detect this in the sled data because this species was poorly sampled. This is likely due to its rather
sparse distribution and large size relative to the mouth
of the sled. It is not surprising that the greatest difference in biomass between reference and trawled corridors occurred in 1995 (Fig. 7) since, by then, the
trawled corridors had been trawled in 3 successive
years. However, it is puzzling that the least difference
occurred in 1994. This could be due to the fact that, in
that year, the reference corridors were sampled after
trawling instead of before (Table 2) and outward
migration of mobile species from the reference corridors into the trawled corridors may have reduced biomass.
120
Mar Ecol Prog SE
The lower biomass of epibenthic invertebrates
observed in the trawled corridors can be due to several
factors which operate over different time scales. Direct
removal by the otter trawl appears to be insignificant
since, as discussed above, the otter trawl is very inefficient in collecting the eplbenthic species common at
the study site. A more likely factor, with a time scale on
the order of hours to days, is predation by scavengers
on dead or damaged organisms left behind on the
seabed. Scavenging by finfish in the wake of trawling
or dredging has been documented by Caddy (1973)
and Kaiser & Spencer (1994) but was probably negligible in our experiment because of the extremely low
abundance of fish (Fig. 6). Due to there being no evidence of any increase in finfish abundance with
increasing trawl set, there is nothing to suggest an
inward migration of fish to feed.
A potential problem that arises with spatially separated treatments is confounding of treatment effects
with location effects (Underwood 1992). However, we
feel that our results reflect true trawling effects and are
not the result of chance differences between the
trawled and reference corridors (300 m apart) in our
13 km long experimental corridors (Fig. 2 ) . Chance
physical effects acting upon the seabed, obscuring
trawling effects, are unlikely since imaging of the
seabed within all corridors was comprehensive both
before and after trawling and sbowed no apparent natural differences between trawled and reference corridors (Schwinghamer et al. 1998). Similarly, no natural
biological processes can be envisioned operating in a
strictly linear fashion in this relatively homogeneous
habitat. For ~nstance,recruitment of bivalves occurred
throughout the entire study area during the 3 yr period
of the experiment (unpubl. data). In addition, grab
sampling conducted prior to the first trawling event in
1993 indicated no significant differences between reference and trawled corridors in total biomass of macrofauna, the biomass of 27 dominant macrofauna species, and size distributions of annelhds, crustaceans and
molluscs (unpubl. data).
Increased attention has recently been given to the
role of invertebrate scavengers (Kaiser & Spencer
1996, Ramsey et al. 1996). The importance of scavengers in marlne ecosystems disturbed by human
activities has been recognised (Britton & Morton 1994).
The snow crab Chionoecetes opilio and sea urchin
Strongylocenfrotus pallidus are well known scavengers on the Grand Banks and responses of these
species to mobile gear can be expected. On one hand,
they can be removed, damaged, or killed by the gear
whlle, on the other hand, those that survive the disturbance, or can migrate into the disturbed area, may
benefit from the increased availability of food items for
scavenging and perhaps from decreased predation by
groundfish because of their removal. This appears to
be the case for snow crabs. The 'functional' response of
snow crab depletion by trawling during the first 6 sets
(Fig. 3) appears to be followed by a behavioural or
'adaptive' response as adult male snow crabs migrated
into the disturbed area during the later sets (Fig. 5).
Direct evtdence of this inward migration by snow crabs
was seen in the video imagery obtained using BRUTIV
(Schwinghamer et al. 1998). The snow crabs collected
in sled samples represent the animals not removed by
the trawl plus those entering the trawled corridor to
scavenge. Therefore, our estimates of the impact of
trawling on snow crab biomass (Tab1.e~
4 & 7) are probably low. The available data indicate a net negative
effect of repeated trawling on snow crabs but not a
simple linear relationship. If sea urchins have a similar
migratory response, we were unable to detect it quant~tativelywith our sampling program. However, video
observations were made of sea urchins apparently
scavenging on dead snow crabs in trawled corridors
(Schtvinghamer et al. 1998).The possibility also exists
that some mobile organisms could migrate out of the
disturbed area.
The ICES Working Group on Ecosystem Effects of
Fishing Activities (ICES 1996) formulated hypotheses
on how the benthic community may change in areas
where trawling is precluded. Closed area studies were
considered as an indirect approach to evaluating the
effects of trawling on seabed habitats. One prediction
was that there would be a shift in size structure toward
large individuals. Our experiment provided direct evidence that otter trawling does decrease the mean individual biomass of epibenthic species (Table 5 ) . However, at the species level, only the sand dollar
Echinarachnius parrna was significantly smaller in
trawled corridors compared to reference corridors
(Table 7 ) . All other speci.es demon.strating significant
biomass reduction had comparatively narrow size
spectra, and immediate effects on their mean individual biomass were not detected. Longer-term effects on
the size composition of sand dol1a.r~could not be evaluated due to the narrowness of the trawled corridors
and possible migration. The trawl bycatch indicates
inward migration of adult male snow crabs during
trawling (Fig. 5) but this did not sign~ficantlyalter the
mean individual biomass of the snow crabs collected
with the sled. This was likely a reflection of the low
catchability of large crabs by the sled, as clearly
observable in video records of sled sampling.
Varying levels of damage on epibenthic and shallow
infaunal bivalves have been reported for different
types of mobile fishing gear (e.g Arntz & Weber 1972,
Caddy 1973, Rumohr & Krost 1991, Shepard & Auster
1991, Bergman & Santbrink 1994, Santbrink &
Bergman 1994, Witbaard & Klein 1994, Kaiser &
Prena et al.: Effects of otter t r a w l ~ n go n benthic epifauna
Spencer 1996, Tuck et al. 1998). Existing theories concerning susceptibility of bivalves to damage are based
primarily on shell mechanical strength relationships. In
general, thin-shelled bivalves are more easily damaged by otter trawl doors than thick-shelled bivalves
(Rumohr & Krost 1991). However, there is conflicting
evidence of the effect of shell size on damage. For
example, Rumohr & Krost (1991)reported that damage
to Arctica islandica increased with size while Witbaard
& Klein (1994) concluded that intermediate sizes were
most susceptible to damage. We observed no significant damage from trawling in the 4 dominant species
of molluscs collected by the epibenthic sled (Tables 7 &
9). These results were surprising at first since the
epibenthic sled samples organisms which live on or
near the sediment surface (i.e. upper 2 to 3 cm) and
therefore are exposed to the destructive forces of the
otter trawl. However, Gilkinson et al. (1998) have
shown experimentally that, on sandy substrata, small
and medium size bivalves living on or near the sediment surface are displaced in fluidized sediment
ahead of the trawl doors and thereby escape direct
hits. Only 2 out of 42 recovered bivalves which had
originally been buried in the scour path of the door
were damaged. These were a large Clinocardium ciliatum at the sediment-water interface and a medium
Macoma calcarea at a depth of 2 cm. It appears that
burial depth combined with size is a key sensitivity
parameter for bivalves.
The undisturbed benthic community at the study site
is relatively homogeneous which was one of the reasons for selecting it (Prena et al. 1996). Regression
analysis of organism abundance data collected with
the sled clearly indicates an increase in the standard
deviation relative to the mean in the trawled corridors
(Fig. 8, Table 10). An increase in the standard deviation relative to the mean was also evident in the trawl
bycatch data (Fig. 3B). This increased variability can
be interpreted as an indication of decreased homogeneity and increased aggregation of the benthic community resulting from stress (Warwick & Clarke 1993).
In this experiment, the stress of trawling could not be
applied uniforn~ly.As a consequence, the level of disturbance in a given sample is dependent on the number of trawl sets and the components of the trawl contacting each individual sample site. The 2 doors, which
are in continuous contact with the seabed, have the
greatest physical impact and affect 2 linear zones
about 1 to 2 m wide. The rockhopper gear and net,
which are in contact with the seabed most of the time
during a set, affect an area about 20 m wide midway
between the doors. The remaining path of the trawl is
influenced by the ground warps which, while not in
direct contact with the seabed, can create turbulence
that resuspends sediment. Another possibility for the
121
decreased homogeneity observed in the trawled corridors is that organisms were redistributed unevenly
after passage of the trawl as suggested by Bergman &
Hup (1992) and Eleftheriou & Robertson (1992).
No attempt was made in this experiment to study the
condition of the invertebrate bycatch once on board
the 'Templeman'. Some individuals may have been
able to survive if immediately dumped overboard. One
of us (K.G.)has observed that in general a high proportion of gastropods in the bycatch of otter trawls on
the Grand Banks appear to be in good condition. Survival experiments of bycatch in beam trawls have been
conducted (van Beek et al. 1990, Fonds et al. 1992,
Kaiser & Spencer 1995). The results indicate that the
mortality of discards is relatively low for starfish
( < l 0%), intermediate for most crustaceans and shellfish, but quite high for small fish ( > g o % ) .The principal
cause of damage in these beam trawl studies appeared
to be the chain matrix and tickler chains fitted to the
gear. These were not present on the otter trawl used in
our experiment. Therefore, it is reasonable to expect
that the survival of bycatch in otter trawls could potentially be higher than in beam trawls; however, proper
experiments are needed to prove this. Demersal fishing may increase whelk mortality in those individuals
that are not captured but which come into contact with
the gear as a result of diminished escape responses
from predators (Ramsey & Kaiser 1998).
Over the longer term, other factors having the potential to affect the biomass of epibenthic organisms in the
trawled corridors include changes in habitat structure
and biogeochemical exchanges between the sediment
and the water column. These could affect the suitability of the seabed as habitat for both adults and younger
life stages. As discussed above, the observed effects of
the experimental trawling on habitat structure seem to
disappear in about 1 yr (Schwinghamer et al. 1998).We
have no information on possible changes in biogeochemical exchanges. One final factor that may also
have influenced the observed difference in epibenthic
biomass between trawled and reference corridors is
changes in the catchability of organisms by the sled
(e.g. burial by resuspended sediment).
The gradual reduction in the biomass of both fish
and invertebrate bycatch in the otter trawl over the 3 yr
experiment is striking (Fig. 6).However, these changes
are not necessarily due to the experimental trawling
but could reflect natural variability or human impacts
over much larger temporal and spatial scales than
included in the design of this experiment. Unfortunately, there are no suitable data bases available for
the general region of the study site that can be used for
comparison. It is well known that there has been a
steady decline in groundfish stocks in recent years on
the Grand Banks. This has brought about the current
122
Mar Ecol Prog Ser
fishing moratorium (Hutchings 1996). This suggests
that the observed drop in fish biomass (Fig. 6) is plausible and probably reflects declines independent of our
experiment. Such a drop in fish biomass should also
reduce predation on benthic organisms which, in turn,
could increase benthic biomass. However, just the
opposite was observed in the invertebrate bycatch by
the otter trawl over the 3 yr experiment (Fig. 6). This
observed decrease is highly suspect. As discussed
above, the otter trawl is very inefficient in collecting
epibenthic organisms. Furthermore, the biomass of
organisms collected by the epibenthic sled (a much
more quantitative device for collecting epibenthic
organisms) in both trawled and reference corridors did
not drop steadily with year (Fig. 7).
Five species were adversely affected by the otter
trawling disturbance applied in our experiment. They
are the snow crab Chionoecetes opilio, sand dollar
Echinarachnius parma, sea urchin Strongylocentrotus
pallidus, brittle star Ophiura sarsi and soft coral
Gersernia sp. (Table 7 ) . These are all relatively large
epibenthic forms that live on the sediment surface and
are therefore most susceptible to capture and damage
from the otter trawl. The 4 dominant species captured
by the sled for which there were no clear evidence of
trawling effects were all molluscs (Astarte borealis,
Margantes sordidus, Clinocardium ciliaturn and Cyclocardia novangliae). These observations are consistent with the results of other impact studies conducted
with otter trawls and beam trawls. In general, the
organisms most affected include epibenthic forms of
molluscs, echinoderms, crustaceans, sponges, hydrozoans, bryozoans, and fish (e.g. Graham 1955, Caddy
1973, de Groot 1984, Rumohr & Krost 1991, Bergman
& Hup 1992, Eleftheriou & Robertson 1992, Kaiser &
Spencer 1994, Auster et al. 1996, Collie et al. 1997,
Jennings & Kaiser 1998). On the other hand, the effects
of trawling are less apparent on infaunal species (e.g.
van Dolah et al. 1991, Bergman & Hup 1992, Brylinsky
et al. 1994, Gilkinson et al. 1998).
In conclusion, this experiment demonstrates that,
along with the obvious ~mmediateimpacts of removal,
mortality and injury on benthic invertebrates seen in
the otter trawl bycatch, other less visible but nevertheless significant impacts on both benthic habitat and
communities can be induced by otter trawling on a
sandy seabed. This is a habitat exposed to storm waves
and iceberg scouring where it would be expected that
the benthic fauna would be already adapted to some
degree of physical disturbance. In particular, by sampling the benthic community with an epibenthic sled,
we were able to detect decreases in total and mean
individual biomass, damage to organisms and decreases in species homogeneity in corridors that were
trawled repeatedly by an Engel 145 otter trawl.
Acknowledgements. First and foremost we thank the crews of
the CSS 'Parizeau' and 'Wilfred Templeman' for their able
assistance in carrying out this major field experimen.t. The
Geological Survey of Canada (Atlantic) provided the Trackpoint and AGCNav systems. We thank IM. Bergman (Netherlands Institute for Sea Research) for her advice on experimental design and modifications to the epibenthic sled. D. Reimer,
D. Harvey, K. Bentham and H. Wiele provided technical support for the sampling equipment. Numerous colleagues
including M. Hawryluk, E. Hartgers, T Mllligan, K. Muschenheim, R. Sweeney, A. Ducharme, K. MacIsaac and K. Saunders assisted in processing the epibenthic sled samples at sea.
P. Woo, E. Hartgers and K. Maclsaac completed the processing of epibenthic sled samples ashore. K. MacIsaac and C.
Bourbonnais assisted in developing the electronic data bases
We thank D. Kulka (DFO) and D. Pitcher (Jacques Whitford)
for their help In estimating the intensity of trawling effort. D.
Pitcher kindly assisted in the preparation of Fig. 1. B. Hargrave, P. Cranford, D. Schneider and 4 referees offered useful
comments w h ~ c hhelped to improve earlier versions of the
manuscript. Funding for this project was provided by DFO ABase, the Northern Cod Science Program, the Atlantic Fisheries Adjustment Program and the Green Plan Sustainable
Fisheries Program. Part~cipationof J.P. was funded by the
Deutsche Forschungsgemeinschaft
LITERATURE CITED
Andrew NL, Pepperell JG (1992)The by-catch of shrimp trawl
fisheries. Oceanogr Mar Biol Annu Rev 30:527-565
Arntz WE, Weber W (1972) Cyprina islandica L (Mollusca,
B~valvia)as food for cod and dab in Kiel Bay. Fish Res
Board Can Trans Series No. 1985. Translated from Bericht
der Deutschen Wissenschafthchen Kommission fiir Meeresforschung (1970) 21:193-209
Auster PJ, Malatesta RJ, Langton RW, Watl.ing L, Val.entine
PC, Donaldson CLS, Langton EW, Shepard AN, Babb IG
(1996)The impacts of mobile fishing gear on seafloor habitats in the Gulf of Maine (Northwest Atlantic): implications
for conservation of fish populations. Rev Fish Sci 4:
185-202
Barrie JV, Le~visCFM, Fader GB, King LH (1984) Seabed processes on the northeastern Grand Banks of Newfoundland; modern reworking of relict sediments. Mar Geol57:
209-227
Bergman MJN, Hup M (1992) Direct effects of beam trawling
on macrofauna in a sandy sediment in the southern North
Sea. ICES J Mar Sci 49:5-11
Bergman MJN, Santbnnk van JW (1.994)Direct effects of beam
trawling on macrofauna in sandy areas off the Dutch coast
In: d e Groot SJ, Lindeboom HJ (eds)Environmental impact
of bottom gears on benthic fauna in relation to natural resource management and protection of the North Sea. N10Z
Rapport 1994-1 1, RIVO-DLO Report C026/94:179-208
Britton JC, Morton B (1994) Marine carrion and scavengers.
Ocednogr Mar Biol Annu Rev 32:369-434
Brylinsky M. Gibson J, Gordon DC J r (1994) Impact of flounder trawls on the intertidal habitat and community of the
Minas B a s ~ n ,Bay of Fundy. Can J Fish Aquat Sci 51:
650-661
Caddy J F (1973) Underwater observations on tracks of
dredges and trawls and some effects of dredging on a scallop ground. J Fish Res Board Can 30:173-180
Collie JS, Escanero CA. Valentine PC (1997) Effects of bottom
fishing on the benthic megafauna of Georges Bank. Mar
Ecol Prog Ser 155159-172
Prena et al.: Effects of otter trawling on benthic epifauna
Dayton PK, Thrush SF, Agardy MT. Hofman RJ (1995) Environmental effects of marine fishing. Aquatic Conservation: Mar Freshwat Ecosyst 5205-232
de Groot SJ (1984) The impact of bottom trawling on benthic
fauna of the North Sea. Ocean Manage 9:177-190
Eleftheriou A, Robertson MR (1992)The effects of experimental scallop dredging on the fauna and physical environment of a shallow sandy community Neth J Sea Res 30:
289-299
Fader GB, Miller R 0 (1986) A reconnaissance study of the
surficial and shallow bedrock geology of the southeastern
Grand Banks of Newfoundland. Geol S U N Can Pap
86-lB, p 591-604
Fonds M, Verboom BL, Groeneveld K (1992) Catch composition and survival of fish and benthic invertebrates in commercial beam trawls for sole and plaice fishery in the
southern North Sea. NIOZ Rapport 16:17-22
Gilkinson K, Paulin M, Hurley S, Schwinghamer P (1998)
Impacts of trawl door scouring on infaunal bivalves:
results of a physical trawl door modelfdense sand interaction. J Exp Mar Biol Ecol 224:291-312
Graham M (1955) Effects of trawling on animals of the seabed. Deep Sea Res (Suppl)3:l-6
Gruffydd LD (1972) Mortality of scallops on a Manx scallop
bed due to fishing. J Mar Biol Assoc UK 52:449-455
Hall SJ (1994) Physical disturbance and marine benthic communities: Me in unconsolidated sediments. Oceanogr Mar
Biol Annu Rev 321179-239
Hutchings JA (1996) Spatial and temporal variation in the
density of northern cod and a review of hypotheses for the
stock's collapse. Can J Fish Aquat Sci 53:943-962
ICES (1996) Report of the Working Group on Ecosystem
Effects of Fishing Activities. ICES CM 1996/Assess/Env:l,
p 126
Jenner KA, Strong KW, Pockhngton P (1991).A review of fishing related seabed disturbance in the Scotia-Fundy
Region Project Report No. 166, Industry and Native Fishenes Branch, Department of Fisheries and Oceans, Halifax
Jennings S, Kaiser MJ (1998) The effects of fishing on marine
ecosystems. Adv Mar Biol34:201-352
Jones JB (1992) Environmental impact of trawling on the sea
bed: a review. NZ J Mar Freshwat Res 26:59-67
Kaiser MJ. Spencer BE (1994) Fish scavenging behaviour in
recently trawled areas. Mar Ecol Prog Ser 112:41-49
Kaiser MJ, Spencer BE (1995) Survival of by-catch from a
beam trawl. Mar Ecol Prog Ser 126:3138
Kaiser MJ. Spencer BE (1996) Behavioural responses of scavengers to beam trawl disturbance. In: Greenstreet SPR,
Tasker ML (eds)Aquatic predators and their prey. Fishing
News Books, Oxford, p 116-123
Krost P, Bernhard M, Werner F, Hukriede W (1990) Otter
trawl tracks in Kiel Bay (Western Baltic) mapped by sidescan sonar. Meeresforsch 32:344-353
Kulka DW (1991) A description of fisheries In the vicinity of
significant oil discovery areas off Newfoundland. Can
Tech Rep Fish Aquat Sci 1787
McCallum BR, Walsh SJ (1997) Groundfish survey trawls
used at the Northwest Atlantic Fisheries Centre, 197l to
present. NAFO Sci Council Studies 29:93-104
McKeo~vnDL, Gordon DC J r (1997) Grand Banks otter trawling impact experiment: 11. Navigation procedures and
results. Can Tech Rep Fish Aquat Sci 2159
McLoughlin RJ, Young PC, )Martin RB, Parslow J (1991) The
Australian scallop dredge: estimates of catching efficiency
and associated indirect fishing mortality. Fish Res 11:l-24
Messieh SN, Rowel1 TW, Peer DL, Cranford PJ (1991) The
123
effects of t r a w h g , dredging and ocean dumping on the
eastern Canadian continental shelf seabed. Cont Shelf Res
11:1237-1263
Meyer TL, Cooper RA, Pecci KJ (1981) The performance and
environmental effects of a hydraulic clam dredge. Mar
Fish Rev 43:14-22
Minitab (1996) Minitab reference manual: Release l 1 for Windows. Minitab Inc, State College PA
Mosteller F, Tukey JW (1977) Data analysis and regression. A
second course in statistics. Addison-Wesley Publishmg
CO,Reading, MA
Prena J , Rowel1 TW, Schwinghamer P, Gllkinson K, Gordon
DC Jr (1996) Grand Banks otter trawling lmpact experiment: I. Site selection process, with a description of macrofauna1 communities. Can Tech Rep Fish Aquat Sci 2094
Raloff J (1996) Fishing for answers. Deep trawls leave
for how long? Science
destruction in their wake-but
News 50:268-271
Ramsey K. Kaiser MJ (1998) Demersal fishing disturbance
increases predation risk for whelks (Buccinwn undatum
L.). J Sea Res 39:299-304
Ramsey K, Kaiser MJ, Hughes RN (1996) Changes in hermit
crab feeding patterns in response to trawling disturbance.
Mar Ecol Prog Ser 144:63-72
Rowell TW, Schwinghamer P, Chin-Yee M, Gilkinson K, Gordon DC Jr, Hartgers E, Hawryluk M, McKeown DL, Prena
J , Reimer D, Sonnichsen G, Steeves G , Vass WP, Vine R,
Woo P (1997) Grand Banks otter trawling impact experiment: 11. Experimental design and sampling equipment.
Can Tech Rep Fish Aquat Sci. 2190
Rumohr H. Krost P (1991) Experimental evidence of damage
to benthos by bottom trawling with special reference to
Arctica islandica. Meeresforsch 33:340-345
Santbrink JW van, Bergman MJN (1994) Direct effects of
beam trawling on macrofauna in a soft bottom area in the
southern North Sea. In: d e Groot SJ, Lindeboom H J (eds)
Environmental impact of bottom gears on benthic fauna in
relation to natural resource management and protection of
the North Sea. NIOZ Rapport 1994-1 1, RIVO-DLO Report
CO26/94:147-138
Schneider DC, Gagnon JM, Gilkinson KD (1987) Patchiness
of epibenthic megafauna on the outer Grand Banks of
Newfoundland. Mar Ecol Prog Ser 3911-13
Schwinghamer P. Guigne JY, Siu WC (1996) Quantifying the
impact of trawling on benthic habitat structure using high
resolution acoustics and chaos theory. Can J Fish Aquat
Sci 53:288-296
Schwinghamer P, Gordon DC Jr, Rowel1 TW. Prena J, McKeown DL. Sonnichsen G, Guigne JY (1998) Effects of
experimental otter trawling on surficial sediment properties of a sandy bottom ecosystem on the Grand Banks of
Newfoundland. Conserv Biol 12:1215-1222
Shepard AN, Auster PJ (1991) Incidental (non-capture) damage to scallops caused by draggmg on rock and sand substrates. In: Shumway SE, Sandifer PA (eds) An international compendium of scallop biology and culture.
Selected papers from the 7th International Pectinid Workshop, World Aquaculture Society, p 219-230
Suuronen P, Lehtonen E, Tschernij V, Larson PO (1996) Skin
injury and mortality of Baltic cod escaping from trawl
codends equipped with exit windows. Arch Fish Mar Res
44:165-178
Systat (1992) Statistics version 5.2. Systat Inc, Evanston, IL
Thouzeau G , V i e R (1991) L'echantillonnage du megabenthos e n zone hauturiere: technique developpee sur le
Georges Bank (Atlantique nord-ouest). CR Acad Sci Paris
312 (Serie III):607-613
124
Mar Ecol Prog Ser 181. 107-124, 1999
Thrush SF, Hewitt JE, Cummings VJ, Dayton PK (1995) The
impact of habitat disturbance by scallop dredging on
marine benthic communities: what can be predicted
from the results of experiments? Mar Ecol Prog Ser 129:
141-150
Tuck ID, Hall SJ, Robertson MR, Armstrong E, Basford DJ
(1998) Effects of physical trawling disturbance in a previously unfished sheltered Scottish sea loch. Mar Ecol Prog
Ser 162:22?-242
Underwood AJ (1992) Beyond BACI: sampling designs that
might reliably detect environmental disturbances. Ecol
Appl4:3-15
van Beek FA, van Leeuwen PI, Rijnsdorp AD (1990) O n the
survival of plaice and sole discards in the otter-trawl and
beam-trawl fisheries in the North Sea. Neth J Sea Res 26:
151-160
van Llolah RF, LVendt PH, Livisen MV (1991) A study of the
effects of shrimp trawling on benthic communities in two
South Carolina sounds. Fisheries Res 12:139-157
Warwick RM, Clarke KR (1993) Increased variability a s a
symptom of stress in marine communities. J Exp Mar Biol
Ecol 172:215-226
Witbaard R, Kle~nR (1994) Long-term trends on the effects of
the southern North Sea beamtrawl fishery on the bivalve
mollusc Arctica islandica L. [Mollusca, Bivalvia). ICES J
Mar Sci 51:99-105
Editorial responsibiLity: David Schneider (Contributing
Editor), St. Johns, Newfoundland, Canada
Submitted: October 27, 1997; Accepted: December 1, 1998
Proofs received from author(s): April 27, 1999